中速磨煤机特征分类及劣化分析算法研究

Research on Feature Classification and Deterioration Analysis Algorithm of Medium-speed Coal Mills

针对Simhash算法特征分类过程存在局部敏感hash、目前中速磨煤机劣化分析算法精准度不高的问题,提出一种中速磨煤机特征分类及劣化分析算法。首先,该算法基于Simhash机制结合余弦定理和矩阵相似性设计特征分类方法;然后,基于八邻域机制结合自适应阈值设计中速磨煤机劣化分析判别规则。结果表明:本文算法使得中速磨煤机的特征参数辨识度、故障关联性以及健康状态预测精度得到了显著提升,且相较于目前主流机器学习算法,在均方根误差(RMSE)损失较小的情况下,计算效率显著提升,所提算法的准确性和健壮性更高。

To address issues of local sensitive hash in the feature classification process of the Simhash algorithm and the low accuracy of deterioration analysis algorithm for medium-speed coal mills,a feature classification and deterioration analysis algorithm of medium-speed coal mills was proposed.Firstly,a feature classification method was designed based on the Simhash mechanism combined with cosine theorem and matrix similarity.Then,the judgment rules for the deterioration analysis of medium-speed coal mills were designed based on the eight-neighborhood mechanism and the adaptive thresholds.Results show that the identification degree of characteristic parameters,the fault correlation and the health state prediction accuracy of medium-speed coal mills are improved significantly.Compared with the current mainstream machine learning algorithms,the calculation efficiency is enhanced significantly when the RMSE loss is small,and the accuracy and robustness of the proposed algorithm are higher.